On The Planning and Operation of Completely Green Microgrids
Abstract
Due to concerns over the rise in global greenhouse emissions from electricity production, an increase in the utilization of renewable energy sources (RES) is becoming imperative in the energy industry. Green microgrids are isolated, small-scale power systems that combine distributed RES and loads into autonomous systems. A completely green microgrid relies exclusively on RES as its energy source. It is expected that green systems, such as green microgrids, will boost the RES usage. Though the planning and operation of microgrids (MGs) have been researched extensively, few current studies exploit MG loads’ characteristics. Accordingly, this research seeks to utilize load characteristics in completely green MGs to: (1) minimize the MG planning and operations costs, (2) characterize the MG performance, and (3) devise time-efficient resource scheduling schemes for MGs. I first considered planning a completely green MG located in a residential community with smart homes. These would contain programmable appliances such as laundry machines and dishwashers, whose operation can be interrupted or shifted in time. The planning problem seeks to determine the optimal number of RES (such as solar panels and wind turbines), as well as the energy storage size that meets the appliances’ load demand in a cost-effective way, while satisfying MG reliability constraints. I use stochastic methods, including Chance Constrained Programming and Monte Carlo Simulation, to account for the randomness in renewable energy production. The study’s numerical analyses show that appliance scheduling can typically reduce MG planning costs by over 40%. Isolated green MGs can also include thermal generators, such as diesel engines and fuel cells, as back-up energy sources to offset unforeseen shortages of renewable energy production. Thus, it becomes crucial to optimally schedule the power generation of these thermal generators to minimize MG operation costs. I exploit the flexibility to schedule programmable appliances in an isolated residential MG to design a time-efficient algorithm that determines a cost-efficient schedule for the thermal generators. The proposed algorithm returns schedules that are very close to those of the optimal or near optimal solutions based on search optimization methods, and with significantly lower time complexity. Finally, I investigate the optimal planning of a completely green charging system for electric vehicles (EVs), which is a completely green MG that supplies energy for EV charging. The study determines the optimal number of solar panels and energy storage capacity that minimizes MG investment costs, while satisfying EV charging performance requirements. I use a three dimensional Markov chain model to account for the intermittency in renewable energy production. Simulation is used to validate the model’s performance results.